AI / ML: Talawa Features Supported by These Technologies

palisadoes

palisadoes

Administrator
The challenges and opportunities of both machine learning (ML) and artificial intelligence (AI) in society have been an ongoing discussion within the Palisadoes Foundation's Talawa GitHub member community. Identifying and justifying efforts to improve Talawa using these technologies has been ad-hoc, so with our new community forum we'd like to ask you for input.

Please let us know how you seriously feel about this.
  1. What new or existing Talawa features could be supported by AI?
  2. What benefit would it provide the Talawa user, administrator or DevOPs engineer?
  3. How likely would they to be adopted?
  4. How can the use of these features be applied independent of a Palisadoes Foundation account with third party services? In other words how would a company implement this on the Talawa feature on own systems without Palisadoes Foundation involvement?
  5. Would the implementation of your proposed feature be a major effort? How could it be implemented?
Please let us know your thoughts.
 
hey @palisadoes

No moderation system exists in posts right now. Pre-trained models could auto-flag problematic content.
AI could:
  • pre flag suspicious content for admin review (not auto-delete)
  • prioritize moderation queue (high-risk content first)
  • suggest similar patterns ("3 other posts from this user were flagged")
  • reduce false reports (filter out spam reports)
Click to expand...
 
  1. Administrators would be interested in analyzing attendance trends and be alerted to people who could be at risk of not attending permanently.
  2. The technology could also be used to detect member status changes and the potential impact on organization events. For example, a wedding may mean events focused on singles may be replaced by events for couples. Appropriate messaging would need to be provided. It may also impact the type of advertising the platform presents.
  3. Advertising could be targeted based on user profiles, their interaction with posts, and events attended.
 
If we use LLMs, for some features like posts and chats
  • Automated Content Moderation: AI models (including LLMs fine-tuned for moderation) can scan forum posts, chats, and event descriptions to flag inappropriate or spam content, reducing manual workload for admins.
  • Event Summary & Knowledge Base: LLMs can automatically generate concise summaries of past events, meeting notes, Q&A from chat discussions, or extract frequently asked questions, making information retrieval fast and contextual.

    Actually with more features coming by we can actually extend more AI usability, RAG System etc.,
 
You must log in or register to reply here.
Back
Top